Martin Holmes wrote:
> > You say that using <note> "amounts to a claim... which is not always
> > going to be true" . Why? Surely it *is* always going to be true? IMT is
> > a tool for linking annotations with images, not transcribing them, isn't
> > it?
> Actually, no; that's been the point all along. In the same document in
> our Mariage project, for instance, we have some divs which contain
> direct transcriptions of text (poems, "speech bubbles" etc.) on the
> engravings, and other divs which have editorial content.
I wondered about that: thank you for the clarification about the
intended scope of application for IMT. Do you think people might use it
to produce a complete transcription of a text though?
IMT is a tool
> for doing whatever users want to do.
Including make life difficult for themselves?
The divs are distinguished by being
> categorized, using a system which is defined in tagsDecl, and which is
> controlled by the user; they can define any categories they like for
> their annotations,
But they *are* annotations then. Yet you would rather allow for one lot
of people to say <div type="note"> and another <div type="annotation">
than accept that these are both the same kind of thing, a thing which
the TEI has already recognized and identified, and for which it proposes
a perfectly usable element <note>? Why stop there? Why not recommend
<tag type=xxx> for everything? :-)
and one that we use in Mariage happens to be
> Transcription. But use of categories is optional, and the definition of
> them is intended to be simple and user-friendly, and not to make
> assumptions about what they might mean.
What does this mean? Defining categories in a simple user-friendly way
is clearly a good thing. If you use the TEI though, you *are* making
assumptions -- or claiming to share assumptions -- about what some of
the categories it defines mean. It's pretty user-unfriendly to use <p>
to mark up pages, or <l> to mark up typographic lines, and I suggest
that it is also pretty unfriendly to use a generic element like <div>
where there is a specific one in existence to do the same job.
> It may seem higgledy-piggledy to you, but you're not an intended user of
> the tool; you'd be quite comfortable figuring out the ulx, uly, lrx and
> lry attributes of your zones manually and doing all your markup in
Actually this accusation is quite false -- I am utterly crap at working
out what the attributes of my zones should be (see earlier drafts of the
PH chapter for evidence) and am looking forward to being able to use a
good tool to sort out such trivia for me. But like other ad hominem
arguments this has nothing to do with the point at issue and I'm
surprised to see you using such an antiquated rhetorical move!
Someone similarly competent with TEI (or even someone like me,
> at the Humpty level) can easily write XSLT to convert the resulting file
> into something more complex and less higgledy-piggledy based on these
> human-defined categories.
Well, but that's the issue though: can they?
> Actually, though, you're answering a question I didn't ask. I just want
> to know about @facs (and/or @corresp). I know very well that most
> advanced TEI-ers are not going to like the IMT markup if they're forced
> to examine it in detail; I just want it to be simple, consistent, valid
> and friendly to novices.
I think all TEI-ers would agree with those goals. Especially the one
I think your suggestion of an embedded note is
> a complication which doesn't add anything, because it's going to be
> wrong as often as it is right, which I think is your objection to plain
My suggestion of <note> adds the explicit semantics that this thing is
an annotation which has been added by someone to the primary text. Since
the goal of IMT is to support the addition of annotations to a facsimile
text (is what you said), I don't see why you think it's going to be
wrong as often as it's right. It isn't *wrong* for a note to include a
transcription of some kind -- especially if you're not usually going to
be transcribing the whole of a text in this way.